TL;DR
This paper introduces a novel convexification approach for the Small-Signal Stability Constrained Optimal Power Flow (SSSC-OPF) problem. By reformulating stability constraints using Bilinear Matrix Inequalities (BMIs) and leveraging Semidefinite Programming (SDP) with objective penalization, the method achieves stable and economical power system operation without relying on traditional, computationally intensive eigenvalue analysis.
What Problem Does This Paper Address?
Ensuring the small-signal stability of power systems is crucial for reliable operation, especially with the increasing integration of intermittent renewable energy sources. The Optimal Power Flow (OPF) problem aims to minimize generation costs while satisfying operational constraints. However, combining economic dispatch with small-signal stability constraints (SSSC-OPF) is challenging because stability conditions are often expressed as nonlinear, nonconvex eigenvalue problems. Traditional methods rely on eigenvalue analysis, which can be computationally expensive, prone to local validity issues due to linearization, and require repeated calculations. This paper tackles the challenge of efficiently solving SSSC-OPF for large-scale power systems by developing a convex formulation that bypasses direct eigenvalue analysis.
Methods
The core of the paper's contribution lies in its novel convexification approach. Instead of directly using eigenvalue-based stability criteria, the authors employ a sufficient condition for small-signal stability expressed as a Bilinear Matrix Inequality (BMI). This BMI is then relaxed into a convex form using Semidefinite Programming (SDP). The formulation also incorporates network structure-preserving Differential Algebraic Equations (DAEs) to model the power system.
Key methodological steps include:
- BMI Relaxation: The nonconvex BMI stability constraint is transformed into a convex form using matrix variable relaxations and SDP techniques. This involves introducing lifting variables and applying techniques like Schur's complement lemma to convert inequalities into Linear Matrix Inequalities (LMIs).
- OPF Relaxation: Standard OPF constraints, particularly the nonconvex power flow equations, are relaxed using lifting variables and semidefinite conditions.
- Stator-Network Coupling Relaxation: Constraints that couple generator dynamics with the network are also relaxed into a convex form.
- Feasible Solution Recovery: To address potential relaxation gaps that might lead to infeasible or unstable solutions, the authors propose an objective penalization strategy. This involves adding penalty terms to the objective function, designed to guide the optimization towards feasible and stable operating points. Vector-norm based penalties are used for large and dense BMI constraints.
Datasets
The proposed method was evaluated on three standard power system test cases:
- WECC 9-bus, 3-machine system
- New England 39-bus, 10-machine system
- IEEE 118-bus, 54-machine system
These are well-established benchmark systems commonly used in power system stability and OPF research.
Key Insights and Ideas
- Eigenvalue-Free Stability: The most significant insight is the successful reformulation of small-signal stability constraints without direct eigenvalue calculations. This bypasses the inherent computational burden and limitations of traditional eigenvalue-based methods.
- BMI to SDP Transformation: The paper demonstrates an effective way to transform complex Bilinear Matrix Inequalities (BMIs) into a solvable Semidefinite Programming (SDP) problem, making stability constraints tractable for large-scale systems.
- Objective Penalization for Feasibility: The proposed objective penalization strategy is crucial for recovering feasible solutions from the convex relaxations. This approach effectively guides the optimizer towards stable operating points, even when the initial relaxed solution might be unstable.
- Trade-off between Cost and Stability: The results highlight the inherent trade-off between generation cost and stability. The proposed method aims to find a stable operating point with minimal additional cost, demonstrating that stability can be achieved without exorbitant economic penalties.
Key Results
The proposed convexified SSSC-OPF (C-SSSC-OPF) method was tested on three benchmark systems and demonstrated the following:
- Achieved Stable Solutions: The method successfully found stable operating points for all tested systems, even in cases where the base relaxed ACOPF solution was unstable.
- Economical Operation: The stable solutions were achieved with minimal additional cost compared to the relaxed ACOPF solutions. For instance, the IEEE 118-bus system achieved a stable solution with only 1.09% additional cost.
- Computational Efficiency: The SDP-based formulation, combined with efficient solvers like MOSEK, makes the approach computationally tractable for large-scale systems, a significant improvement over traditional methods.
- Reduced Error Metrics: The error metrics for relaxation gaps (e.g., $\epsilon_{W}$, $\epsilon_{Wdq}$, $\epsilon_{P}$) were shown to be small, indicating the effectiveness of the convexification and penalization techniques.
Limitations and Future Work
- Global Optimality Guarantee: The paper acknowledges that the objective penalization approach, while effective for recovering feasible solutions, does not guarantee global optimality due to the introduction of a potential optimality gap.
- Tighter Convex Envelopes: The authors suggest that future work could focus on developing tighter convex envelopes for relaxations, particularly for Park's transformation, to further reduce relaxation gaps and improve accuracy.
- Weight Selection: The selection of weights for the multi-objective penalization functions is currently based on estimation and iterative refinement. Developing a more systematic approach for weight selection would be beneficial.
- AVR and PSS Integration: While the Automatic Voltage Regulator (AVR) was modeled, Power System Stabilizers (PSS) were not explicitly considered for testing the method's stability enforcement capability. Future work could explore their integration.
Summary updated